[Intel-gfx] [DMC_REDESIGN_V2 01/14] drm/i915/gen9: csr_init after runtime pm enable

[Imre Deak]

On ke, 2015-08-26 at 16:58 +0530, Animesh Manna wrote:
> Skl is fully dependent on dmc for going to low power state (dc5/dc6).
> This requires a trigger from rpm. To ensure the dmc firmware
> is available for runtime pm support rpm-reference-count is used
> by not releasing the rpm reference if firmware loading is
> not completed.

The above doesn't explain to me why we need this change. Looking at the
next patch makes it clearer: we need to move the firmware loading later
since it needs to take a power domain reference instead of an RPM
reference, and power domains are not initialized at this point yet. It's
also worth mentioning in the commit log that this change is needed by an
upcoming patch.

> 
> So moved the intel_csr_ucode_init call after runtime pm enable.
> 
> Cc: Daniel Vetter <daniel.vetter at intel.com>
> Cc: Damien Lespiau <damien.lespiau at intel.com>
> Cc: Imre Deak <imre.deak at intel.com>
> Cc: Sunil Kamath <sunil.kamath at intel.com>
> Signed-off-by: Animesh Manna <animesh.manna at intel.com>
> ---
>  drivers/gpu/drm/i915/i915_dma.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
> index 2193cc2..48b9792 100644
> --- a/drivers/gpu/drm/i915/i915_dma.c
> +++ b/drivers/gpu/drm/i915/i915_dma.c
> @@ -885,9 +885,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
>  
>  	intel_uncore_init(dev);
>  
> -	/* Load CSR Firmware for SKL */
> -	intel_csr_ucode_init(dev);
> -
>  	ret = i915_gem_gtt_init(dev);
>  	if (ret)
>  		goto out_freecsr;
> @@ -1035,6 +1032,9 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
>  
>  	i915_audio_component_init(dev_priv);
>  
> +	/* Load CSR Firmware for SKL */
> +	intel_csr_ucode_init(dev);

We need to call this function earlier, since there could be a modeset
before this for the console, after which the driver will disable any
unneeded power domains. So this should be called after
intel_power_domains_init_hw() to prevent this.

> +
>  	return 0;
>  
>  out_power_well: